Recently, the development of a storage system that integrates and packages hardware such as calculation nodes (server systems), I/O nodes (storage apparatuses), and network devices has been being promoted. This storage system is also called a converged storage system and is configured in such a manner that the calculation nodes, the I/O nodes, and the network devices are connected respectively in one chassis in advance. Various verifications and settings of the converged storage system are performed in advance in order to optimize the performance of the storage system in the chassis and then the converged storage system is supplied to a user company.
So, the user company to which the converged storage system is supplied can avoid the trouble of having to perform the various verifications and settings by themselves by not ordering the delivery of the calculation nodes, the I/O nodes, and the network devices individually and can start operating the delivered storage system in a short period of time. Moreover, an administrator of this converged storage system can centrally manage each of the calculation nodes, the I/O nodes, and the network devices included in the storage system by using one management device. Accordingly, it is unnecessary to provide a server administrator, a storage administrator, and a network administrator individually and it is possible to significantly reduce operational costs.
In this way, the converged storage system can realize enhancement of the system performance, enhancement of operability, and cost reduction. On the other hand, since it is a system which is packaged in advance, there is a problem of poor flexibility in its hardware configuration.
For example, if an attempt is made to migrate one storage system to another storage system, data in a chassis will have to be transferred to outside the chassis. Under this circumstance, the hardware configuration cannot be changed flexibly, for example, to transfer the data at high speeds by providing a dedicated connecting path between a migration source I/O node and a migration destination I/O node. Therefore, in this case, it is necessary to transfer the data via a general connecting path with a small bandwidth which is unfit for transfer of large amounts of data.
When the large amounts of data are to be transferred via the general connecting path with a small bandwidth, it requires a long time to complete the transfer processing. If the transfer processing is executed continuously for a long period of time, loads are focused on the I/O nodes in the migration source storage system during that time and I/O processing and communication processing are delayed. As a result, this causes a problem of performance degradation of the entire storage system for a long time.
Incidentally, examples of cases where it is required to transfer data in a chassis to outside the chassis include: a case where a storage system currently in use is about to reach the end of its life due to aged deterioration; a case where load leveling processing is executed between the chassis; or a case where power saving processing is executed to power off a chassis which has become surplus as a result of using only one of the chassis for a virtual server.
Accordingly, when data in the chassis of this converged storage system is to be transferred to outside the chassis, it is required to complete the transfer processing in a short period of time. A technique to transfer the data in the chassis to outside the chassis, that is, a technique to transfer the data between I/O nodes is disclosed in, for example, Patent Literature 1 and 2 mentioned below.
Patent Literature 1 discloses a technique to offload the processing for transferring data between I/O nodes to a calculation node. A calculation node in which applications operate is decided as an offload location for the transfer processing. The reason is to ensure consistency between data of a migration source I/O node and data of a migration destination I/O node.
Furthermore, Patent Literature 2 discloses a technique to back up data of a host computer between the host computer and an I/O subsystem, which cannot be connected directly because of the difference in their access interfaces, by transferring the data of the host computer to the I/O subsystem. Like Patent Literature 1, the I/O subsystem which intermediates I/O processing is decided as an offload location for the transfer processing.